Existing Augmented Reality (“AR”) imaging systems are limited in their abilities to recognize objects or render AR elements associated with the objects at precise locations when the objects are not fully captured by a camera. Many of the existing systems recognize objects by analyzing the image of the object. In such system, when the object is blocked from view or when some key features of the object are lacking in the object's image, the system cannot recognize it, let alone rendering the desired AR element(s) of the object at a designated location. As such, such systems cannot be used to render AR elements of some remote objects, such as a stage in a concert where the audience in the far back don't always have a clear view of the entire stage.
Furthermore, for AR elements that are displayed as a frame or a background of an object, existing AR systems are limited in rendering such AR elements based on the predetermined shapes and dimensions of the AR elements or the edge(s) of the object. Frame-like AR elements of fixed shapes and dimensions are boring since they are merely superimposed on the image of the object. Even if the system can detect the edge(s) of the object, such determination is generally based on image processing techniques. Again, once the object is not fully captured by the camera, the edges) of the object cannot be determined, and the AR system would not be able to properly render this kind of AR element(s).
Therefore, it is desired to have a method or system configured to identify one or more objects even when some of the objects are entirely or partially hidden from view.
It is also desired to have a method or system configured to render an AR element with a cut-out/see-through area that is determined based on the characteristics of an object, rather than the properties of the AR element.